Font Size: a A A

Study On The Key Techniques Of Underwater Multi-target Tracking In Guidance System

Posted on:2005-11-07Degree:DoctorType:Dissertation
Country:ChinaCandidate:J W DangFull Text:PDF
GTID:1118360155977389Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
The theory and methods of underwater multi-target tracking in guidance system involves underwater acoustic engineering, signal processing, modern control theory, precise guidance theory, and so on. The study on these theories and methods is very important in many fields. In this dissertation, a multi-target tracking frame based on underwater guidance system is constructed, and some key theories and methods are studied such as the modeling of underwater maneuvering targets, the tracking filter and prediction algorithms, the data association algorithms, multi-target tracking algorithm in different coordinate systems, tracking gate methods and so on. The main achievements and creative results in this dissertation are summarized as follows:1. Based on the study of maneuvering target models, the adaptive Gaussian model for underwater maneuvering targets and the tracking algorithm using this model are presented. The theoretical analysis and computer simulation of this model is accomplished. The results show that the adaptive Gaussian model can adapt the change of target maneuvering acceleration and its variance, and maneuvering detection will not be necessary in this tracking algorithm. So the problem of time delay in maneuvering targets tracking would be eliminated.2. Filter and prediction algorithms for underwater targets tracking are studied. The basic character of discrete Kalman-filtering algorithm and bias error of extended discrete Kalman-filtering algorithm are analyzed. A new Kalman-filtering algorithm of weighting on the square deviation matrix of measure noise is proposed and some simulations are accomplished. The results show that this algorithm has robust ability to the measure error and model error. Even if tracking system has model error and time-vary measure error, this algorithm also can get preferable tracking filtering precision.3. The "nearest neighbor" data association algorithm, probabilistic data association algorithm, interacting multiple model probabilistic data association algorithm and joint probabilistic data association algorithm are studied. The characteristics and suitable circumstance of each algorithm are summarized. The results show that the joint probabilistic data association algorithm is a favorable algorithm in tracking underwater multi-target, yet this algorithm would have a large calculating burden with the increase of targets tracked and return measure.4. Considering the characteristics of lower velocity and acceleration of underwater motion targets, complex underwater acoustic channel, lower data rate and signal-to-noise ratio of sonar system, a new decoupled underwater targets tracking algorithm based on adaptive Gaussian model in spherical coordinate system is proposed. The results show that this algorithm can preferably adapted to underwater targets tracking without any prior knowledge about the maneuvering characteristics of targets tracked. Compared with the conventional tracking algorithm in spherical coordinate system, this algorithm has better tracking performance for target range, azimuth, and elevation.5. A new acceleration weighting adaptive underwater targets tracking algorithm, and a new extended measure underwater targets tracking algorithm in spherical coordinate system is proposed respectively. The results show that because of stressing the new information of targets maneuvering acceleration in the acceleration weighting adaptive tracking algorithm, the motion model of targets tracked can be modified on line. Using the radial velocity measurement of targets tracked in extended measure tracking algorithm in spherical coordinate system, the tracking precision of target range can be improved, meanwhile, the tracking precision of azimuth and elevation can be improved, and the dynamic convergence of the tracking system can also be quickened.6. Considering greater angel measure error and states converting error from spherical coordinate system to cartesian coordinate system in underwater target tracking systems, a novel debiased converted measurement underwater targets tracking algorithm and a novel unbiased converted measurement underwater targets tracking algorithm in cartesian coordinate system is presented respectively, and the tracking errors of these algorithms are analyzed. The results show that the debiased converted measure tracking algorithm is more precise and stable than conventional converted measure tracking algorithm in cartesian coordinate system, and unbiased converted measure tracking algorithm has better adaptive performance to the different probabilistic distribution of angel measure error.7. Based on the study on the Tracking gate algorithms, locating principles of multi-beam system and its implementation, a new underwater multi-target tracking algorithm based on multi-beam locating system is proposed, and some simulations are accomplished. The results show that this tracking algorithm can perform the tracking for the different maneuvering states of targets tracked in a certain extent, but, because of the disadvantage of multi-beam system, there is a greater tracking error as a target maneuvering between different locating beams, and it is hard to distinguish two or more targets in one locating beam.8. Combining the study on the application of high-resolution targets parameter estimation techniques in underwater guidance system, a new underwater multi-target tracking algorithm based on high-resolution parameter estimation is proposed, and some simulations are accomplished. The results show that this tracking algorithm has well tracking performance for underwater multi-target. Even if targets tracks crossing in tracking space, this tracking algorithm has still a high associating success rate and tracking precision.
Keywords/Search Tags:Underwater multi-target tracking, Maneuvering target model, Nonlinear filtering, Coordinate system, Data association, Tracking gate, High-resolution parameter estimation, Tracking algorithm
PDF Full Text Request
Related items